The Emancipation Proclamation, Confederate expectations, and the price of Southern bank notes.

1. Introduction

In 1862, the Northern war effort floundered in the East. Stonewall Jackson threatened Washington, McClellan's peninsular campaign stalled, and the North suffered its second defeat at Bull Run. The end of the conflict and restoration of the Union were nowhere in sight. By the summer, Lincoln concluded: "Those enemies must understand that they cannot experiment for ten years trying to destroy the government, and if they fail still come back into the Union unhurt. (1) Lincoln conveyed this message to Southerners through the Emancipation Proclamation, which effectively eliminated the prospect of a negotiated settlement and a return to the Union unhurt. Was Lincoln's assessment of Southern expectations correct? Did the Emancipation Proclamation alter these expectations as he had anticipated? We present evidence from the market for Southern bank notes that he was right on both counts.

Before the Civil War, the Federal Treasury minted coins, but the United States did not have a uniform currency. State legislatures authorized private banks to issue bank notes redeemable in gold. These bank notes constituted the primary circulating medium of exchange. The Confederacy added its own currency to the circulating bank notes so that it could generate revenue through money creation. This seignorage required the public's acceptance of Confederate currency. To gain this acceptance, the Confederacy neither monopolized currency production nor made Confederate currency legal tender. Instead, Southern states passed legislation that effectively monetized Confederate currency. Thus, Confederate currency and privately issued bank notes circulated side by side.

We argue that the fate of the Confederate currency depended on Northern war aims. Lincoln's initial goal was to preserve the Union; if the preservation of slavery would help achieve this goal, then so be it. (2) However, as the war carried on, Lincoln's objectives changed. The Emancipation Proclamation of September 22, 1862, made these new objectives clear. The Proclamation would take effect on January 1, 1863, and it would radically change the stakes of the war. Lincoln remarked to a member of his administration: "the [old] South is to be destroyed and replaced by new propositions and ideas." (3) None of this was lost on Southerners. Reunification now meant the destruction of their way of life.

In this article, we use a newly constructed semimonthly series of Southern bank note prices to investigate how Northern policy affected these prices in the South. In particular, bank notes traded at par with Confederate currency until November of 1862, when note prices appreciated and then fluctuated. (4) We argue that it was the Northern policy commitment created by the Emancipation Proclamation and validated by the November elections that caused Southern bank note prices to diverge from par with Confederate currency. We also examine the effects of war events on state bank note prices after the divergence occurs. (5)

In the next section, we explain how the Confederate states adopted regulations that monetized Confederate currency. In section 3, we construct a simple illustrative model of the demand for two currencies to show how the relative price of the two depends on expectations of the future value of each of the currencies. In section 4, we discuss how Northern occupation policies and the Emancipation Proclamation influenced Southern expectations. Section 5 presents our data on Southern bank note prices and a discussion of the behavior of these prices during the war. In section 6, we regress the growth rate of the price of North Carolina bank notes on war events and analyze the results. Section 7 concludes.

2. The Establishment of Confederate Currency as a Medium of Exchange

At the outbreak of war, the South expected to finance a short-lived conflict with bond issues serviced by tariffs from cotton exports. The North's effective naval blockade and the length of the war changed these plans. The South was unable to raise sufficient funds through tariff collections, taxation, or bond financing and so turned to money creation. (6) Reliance on seigniorage required the South to generate a demand for Confederate currency as a medium of exchange. Changes in state banking laws were necessary to make this transformation.

Banking systems across the South had diverse structures. Of the 11 seceding states, Texas and Arkansas prohibited banking and Mississippi had few banks. North and South Carolina allowed state chartered banks; Alabama, Florida, Georgia, Louisiana, Tennessee, and Virginia had free banking. (7) Under most free-banking laws, banks were required to purchase and hold state bonds as security against their notes and to exchange their notes for specie on demand and at par. Note holders held the first lien on the assets of a failed bank, and stockholders were personally liable for the losses of the bank up to the value of their investments. (8) The legal obligations of bankers under chartered banking were similar but less clear. In general, there were penalties for failure to redeem notes for specie at par and restrictions on the amount of notes that could be issued. The various bank notes circulated side by side, providing the antebellum South numerous mediums of exchange.

The crisis of 1860 provoked significant political and financial uncertainty that led note holders to redeem bank notes for specie. Gold drained from the banking system, threatening its viability. In order to preserve the specie in the banking system, Southern state legislatures, except Louisiana and Florida, authorized their banks to suspend specie payments before the end of 1860. (9) The gold drain ceased, and bank notes began to trade below face value. (10) By the end of the next year, Louisiana and Florida had followed suit. (11)

The salvation of the system by suspending gold redemption came at a price to banks. The states with banking institutions required banks to accept Confederate currency on deposit at par as a new condition for the suspension of specie payment. (12) Banks across the South also agreed to accept Confederate currency to settle interbank accounts. The Governor of South Carolina, F. W. Pickens, announced, "The banks have agreed to take Treasury Notes in adjustment of all balances between themselves. To this extent, they are thus made equivalent to gold and silver, and of course it gives them almost exclusive circulation." Pickens later hinted how these agreements were secured: "The legislature wisely legalized the suspension of all our banks. It would be proper that you inquire into the conduct of these institutions, and if a course has been pursued by any of them deemed not patriotic or proper, then the benefits of this act should be suspended...." (13)

In addition, Confederate currency was deemed receivable at par for state taxes and public dues. (14) Tax codes before the war resembled the Alabama 1853 statute that required all taxes to be paid in gold and silver, warrants on the state treasury, bills of specie-paying banks of the state of Alabama, or the Treasury notes of the United States. The statute was amended in early 1861 to read "Tax collections of this state may take in payment of taxes: Treasury notes of this state, or of the Confederate states and any current Bank-notes of the state of Alabama...." (15)

Confederate currency became accepted as a medium of exchange because they were received as payment for state dues and taxes, and accepted by banks on deposit, and used in interbank settlements. So Confederate currency circulated even though they were not fully backed by gold. Moreover, the notes of banks that had suspended redemption circulated on the same footing as the Confederate currency. Thus, state legislation effectively created a new medium of exchange that circulated in the South along with bank notes. (16)

3. A Simple Model of Southern Bank Note Prices

The South successfully introduced a Confederate currency that circulated alongside bank notes. We construct a simple model of the demand for two monies to capture the fundamental relationship between Confederate currency and Southern bank notes. (17) We show that, as long as people expected the two monies to continue to trade at par in the future, Confederate currency and bank notes would trade at par in the current period. (18) However, once there was some chance that Confederate currency would lose its status as a medium of exchange, its expected future value fell, and so the bank notes sold at a premium relative to Confederate currency.

We assume that households have a two-period planning horizon and receive "liquidity services" from bank notes and Confederate dollars. These liquidity services can be represented by a money-in-the-utility-function formulation. (19) The various states agreed to accept bank notes and Confederate currency at par in payment of taxes. This provision, along with the banks' use of currency to settle interbank accounts, equalized the current-period liquidity value of bank notes and Confederate currency. This equalization implies that Confederate currency and bank notes enter as a sum in the first-period utility function. We do not impose this restriction on the future liquidity services of the two. The utility function of a typical Southern household is thus

U = u([c.sub.t], [d.sub.t]/[P.sub.t] + [n.sub.t]/[P.sub.t]) + v([c.sub.t+1], [d.sub.t+1]/[P.sub.t+1], [n.sub.t+1]/[P.sub.t+1],

where [c.sub.t] is consumption, [d.sub.t] and [n.sub.t] are the face values of Confederate currency and bank notes, respectively, and [P.sub.t] is the price level, all in time t.

We assume that the household in question receives real income, [y.sub.t], in period t, begins the first period with given quantities of Confederate currency and bank notes, and faces the interest rate R. The household assumes that the second period is its last. The intertemporal budget constraint is thus

[P.sub.t][y.sub.t] (1 + R) + [P.sub.t+1][y.sub.t+1] - [Rd.sub.t] - R[P.sup.N.sub.t][n.sub.t] - ([P.sup.N.sub.t] - [P.sup.N.sub.t+1])[n.sub.t] = [P.sub.t][c.sub.t] (1 + R) + [P.sub.t+1][c.sub.t+1] + [d.sub.t+1] + [P.sup.N.sub.t+1][n.sub.t+1]

where [P.sup.N.sub.t] is the Confederate dollar price of bank notes in period t and [P.sup.N.sub.t+1] is the expected Confederate dollar price of a bank note in period t + 1. The first two terms on the left-hand side represent the future value of the household's income stream, the terms [Rd.sub.t] and R[P.sup.N.sub.t][n.sub.t] represent the interest opportunity cost of holding money, and the term ([P.sup.N.sub.t] - [P.sup.N.sub.t+1]) [n.sub.t] represents the expected capital gain or loss from holding bank notes. The first two terms on the right-hand side are the future value of the consumption stream. The remaining terms represent money holdings in the final period.

The household chooses current and future consumption and money holdings to maximize utility subject to the intertemporal budget constraint. Let [u.sub.2] be the partial derivative of u(*) with respect to its second argument, so that [u.sub.2] = [differential]u/[differential]([d.sub.t]/[P.sub.t]) = [differential]u/[differential]([n.sub.t]/[P.sub.t]). The relevant first-order conditions for our concerns are

[u.sub.2] - [lambda]R = 0 and [u.sub.2] - [lambda](R[P.sup.N.sub.t] + [P.sup.n.sub.t] - [P.sup.n.sub.t+1] = 0,

where [lambda] is a Lagrangian multiplier. (20) Equating the two and eliminating the multiplier yields

[P.sup.n.sub.t]R + ([P.sup.N.sub.t] - [P.sup.n.sub.t+1]) = R.

The perfect substitutability of the two as mediums of exchange implies that the marginal benefit from holding another unit of either note is exactly the same. Optimality thus requires that the marginal cost of holding an additional unit of each also be the same. The left-hand side of the above equation is simply the opportunity cost, measured in Confederate currency, of holding a bank note. This cost is the interest opportunity cost of holding a bank note tempered by the expected capital gain if the relative price of bank notes should rise. The right-hand side is simply the interest opportunity cost of holding a Confederate note.

We can rewrite this condition as

[P.sup.N.sub.t] = 1 + ([P.sup.N.sub.t] - [P.sup.N.sub.t+1]) = R

The value of a bank note is equal to one plus any expected capital gain or loss scaled by the inverse of the interest rate. The role of expectations in the determination of the Confederate dollar price of bank notes is now clear. Any expected appreciation or depreciation of bank notes is reflected in their current price. In the absence of such expectations, bank notes and Confederate currency trade at par. Indeed, given the liquidity equivalence, if bank notes currently trade at par, then it must be the case that the household expects them to continue to trade at par in the next period. Empirically this result means that, if we observe [P.sup.N.sub.t] = 1, then people expect [P.sup.N.sub.t+1] to also equal one. On the other hand, if we observe bank notes selling at a premium, that is, [p.sup.N.sub.t] > 1, then people must expect additional appreciation of bank notes in the future and so expect [P.sup.N.sub.t+1] to exceed one. It is this relationship between the current price of notes and expectations that allows us to extract information on expectations from the market for bank notes.

4. The Evolution of Northern Policy Toward Confederate Money

Union policy toward currency and banking developed as the war progressed. The occupying generals in the field determined Union policies regarding the local Southern currency. The Union captured the rich banking center in New Orleans and some of the coastal branches of North Carolina banks in the early part of 1862. In the face of advancing Union forces, the Confederates shipped most of the gold out of New Orleans before the city was captured in April; but the banks still owned their specie, and their notes were still "backed" by their expected redemption values. General Butler, the Union commander of occupied New Orleans, prohibited the circulation of Confederate currency and held the New Orleans banks liable for all of their outstanding notes, including the notes recently issued in exchange for Confederate currency. (21) Moreover, banks were obliged to redeem their notes in either gold or Northern greenbacks, although this redemption did not have to take place until after the war. Consequently, these New Orleans bank notes circulated and constituted most of the money supply in New Orleans. (22)

The upper coast of North Carolina was occupied from May 1862 until 1864 when Confederate forces recaptured the area. During that time, the branch of the Bank of Cape Fear at Washington, North Carolina, and the branch of the Bank of North Carolina at Windsor, North Carolina, operated under Union occupation. (23) These banks corresponded with their parent banks in Confederate-held territory and continued operation. The small amount of Confederate currency deposited on account was shipped to Confederate territory "for safe keeping," and the coastal branch conducted business solely in bank notes, which at that time exchanged at or near par with Confederate dollars. (24)

Southerners residing inside Confederate-held territory became aware of General Butler's policies and altered their currency portfolios when directly threatened by Union forces. When Charleston came under naval siege, the naval provost reported to the Confederate Secretary of the Treasury C. G. Memminger on June 16, 1862, that "There are many instances of persons coming from Augusta and other places purchasing the notes of our state banks at high premiums as much as ten per cent being sometimes paid. Many of our own people collect bank notes and refuse to part with them, while some it is said even refuse to take Confederate notes in payment." He continued, "One explanation I have heard is that in New Orleans the notes of the state banks were allowed to circulate while Confederate notes were not." (25)

Southerners had thus learned that, in territory occupied by the Northern army, Confederate currency did not circulate. Although occupation policies during hostilities am not policy commitments, they provide useful information that influence expectations about postwar policy.

5. The Confederate Dollar Price of State Bank Notes

Bank Note Prices to November 1862

Our first report of the Confederate dollar price of state bank notes is from the October 1, 1861, Richmond Enquirer. Bank note prices were typically reported by state and not by individual banks in the state. (26) Reports of bank notes trading at par with Confederate currency are made fairly often in the Enquirer and Examiner through October of 1862. Exceptions are sometimes noted for individual banks that faced liquidity problems. For example, the Bank of Pittsylvania apparently experienced some difficulties, and the December 20, 1861, Richmond Examiner reported that the notes from this bank were not taken on deposit. (27)

Even after Antietam, where Lee's invasion was stalled in September of 1862, bank notes continued to trade at par. The South, though outnumbered and with considerably less manufacturing capacity than the North, apparently expected no worse than a negotiated settlement that would leave their social structure intact and their territory unoccupied. Hope was fired by recollections of English troops withdrawing from the colonies, the antiwar movement in the North, and the possibility that England or France might recognize the Confederacy. (28)

Until November of 1862, bank notes continued to trade at par. The calm in the note market was broken on November 5, 1862, when the Richmond Daily Whig reported that bank notes traded at $1.25 Confederate dollars. (29) The path of bank note prices plotted in Figure 1 clearly shows the break in prices and their irregular behavior thereafter. Theory suggests that the sudden departure of bank notes from par was caused by a change in expectations. What explains this sudden change in expectations that had been held by Southerners for over a year?

Prior to the Emancipation Proclamation, the Northern war aim consisted solely in preserving the union. This meant that, even if the war went badly for the South, a negotiated settlement restoring the status quo ante loomed as the worst-case scenario for the seceding states. Both slavery and the state banking systems would have been maintained according to antebellum law. Recall that Confederate currency was never made legal tender by the Confederate Congress. Instead, the Southern states had acted in accordance with their prerogatives over state banks and state tax laws to support the Confederate currency. The returning Southern states would have retained control over their banking institutions and a strong case could have been made for the legality of Confederate currency in the courts. Moreover, a negotiated Southern defeat would have lifted the federal blockade and restored the fiscal capacity of the Southern states to support both Confederate currency and bank notes.

The Emancipation Proclamation changed everything. The Proclamation, validated by the November 1862 elections, committed the North to restructuring Southern society. This virtually eliminated the prospect of a negotiated settlement. By implication, Lincoln embraced the Union occupation policies repudiating Confederate currency and allowing bank notes to circulate. In order to conquer and restructure the South, the whole Confederacy would have to be subjected to occupation policies. For the South and Confederate currency holders, only two possible outcomes remained: Southern independence or Northern rule, and the latter meant that Confederate money would become worthless.

Between the issuance of the Proclamation and the date it was to take effect stood the elections of 1862. The dates of the 1862 Congressional elections varied by state. Elections were held in Indiana, Iowa, Ohio, and Pennsylvania on October 14th. In these elections, Democrats gained seats. To gain control of Congress, the Democrats still needed to gain seats in Massachusetts, Michigan, Minnesota, Missouri, New Jersey, New York, and Wisconsin in the November 4th elections. The Democrats gained some seats, but Republicans and allied parties in support of the Proclamation retained control of the House. The narrow victory did not dissuade Republicans from supporting the Proclamation. (30) On November 5, the New York Herald reported:

   ... the result of the election in this State--following upon that
   of the elections in Ohio, Indiana and Pennsylvania--has given a
   deadly blow to the obstructionists and radicals who would postpone
   the termination of the war to an indefinite period for their own
   base ends.

Thus, we argue that, prior to the Emancipation Proclamation, Southerners expected that the conflict would end with their social structure intact. Given this expectation, it was reasonable for Southerners to expect that Confederate currency would continue to circulate after the war, and this led bank notes and Confederate currency to trade at par. After the Proclamation and its validation by the elections of 1862, the North was committed to total victory and an upheaval of the Southern social system. It was now reasonable to expect that, in the event of a Northern victory, the South would be occupied, an expectation that would be realized. Under wartime occupation, Northern commanders had repudiated Confederate currency. After the elections of 1862, there was some chance, perhaps a good chance, that in the event of a Northern victory, Confederate currency would not be allowed to circulate. We argue that it was this expectation that caused bank note prices to diverge from par in November of 1862.

The timing of the elections and the timing of the departure of bank note prices from par strongly suggests that the Proclamation and its validation by the Congressional elections were the key considerations. However, given the nature of the data and possible lags in information flows, we cannot be sure. An alternative explanation for the departure from par is the Union repulsion of the Confederate army at Antietam. Weidenmier (2002), using a carefully constructed weekly series of gold prices, located a turning point in the Confederate dollar price of gold on October 21, 1862. He argued that the Battle of Antietam was the likely cause of this turning point. The month between Antietam and this price change raises questions about this assignment of cause. Weidenmier recognized this shortcoming and suggested that the lag may have been due to Confederate censorship of war news.

The first wave of elections in the North took place on October 14 and may also explain the October turning point discovered by Weidenmier. Throughout September and until October 17, the price of gold was steady at $2.40. It jumped to $2.80 on October 24, an increase of over 16% in just a week. In addition, the price jumped from $2.80 to $3.30, a near 18% increase, over the next two weeks as the November elections were resolved. So, from October 17 to November 7, the price of gold jumped from $2.40 to $3.30, a nearly 40% increase over the three weeks during which the control of congress was settled. However, the cause of these price changes are further confounded by the passage of the first funding act on October 13, 1862. This act set a deadline for the conversion of Confederate currency into 8% bonds and would have had the effect of raising the price of gold. No definitive answer may ever be given because asset prices may have been, and likely were, affected by all four events: Antietam, the elections in the North, the funding act, and the Emancipation Proclamation.

6. State Bank Note Prices: Empirical Results

After the tight link between bank note prices and Confederate currency was broken, the relative values of Confederate currency and bank notes would, according to our argument, depend on expectations concerning the fate of the Confederacy. Figure 1 suggests that this was so. After November 1862, bank note prices gradually rose to about $2.00 by the end of June. The price of bank notes clearly reflect the Southern battlefield catastrophes in the summer of 1863. By the beginning of August, bank note prices jumped to over $3.00; market reports then vanished until September. When quotes resumed, state bank note prices were relatively stable through the fall of 1863 at about $3.00. In December, following the Union win at Chattanooga, bank note prices rose to $3.25. The Confederate dollar appreciated against state bank notes as the spring campaign of Grant stalled. The appreciation was greater against Virginia notes because Grant's move south toward Richmond directly threatened bank assets there. The fall of Atlanta in August of 1864 is also clearly reflected in note prices.

To check the impression left by Figure 1, we turn to a quantitative analysis of the effect of war events on the Confederate dollar price of state bank notes. We run an ordinary least squares regression of the growth rate of the Confederate dollar price of North Carolina bank notes on a constant and dummy variables for specific Northern and Southern victories. The dummies are equal to one on the date just after the victory and zero elsewhere. (31)

In the newspapers from which we collected our data, sometimes a single price is reported and sometimes a spread is given. Sometimes note prices are identified as bid prices, sometimes they are identified as ask prices, and sometimes no such identification is given. When two prices are reported, we use the average. When only one price is given, we take that price as our observation and do not make any attempt to adjust it if it is known to be a bid or ask price. These conventions introduce unavoidable error into our series of bank note prices. Fortunately, for our application, the error is in the dependent variable and does not present serious estimation difficulties. The sample is semimonthly; it begins in November of 1862 and continues through February of 1865. We took the observations closest to the 7th and 22nd of the month. Growth rates are daily average rates to account for the differing number of days between observations. (32)

The dates of important battles are those McCandless (1996) collected from Civil War military histories by Pratt (1948) and Catton (1971). (33) The dummy variable for Stones River is three days after its conventional date. This lag is the shortest time between the end of the battle and the observation on bank note prices. Below we discuss a possible complication of this timing. Most of the other time spans between the battle and the bank note observation are from five to ten days. The longest span is 19 days for Gettysburg. We chose this long span, as opposed to the alternative, which would have been only two days, because the news from Gettysburg took time to arrive in Richmond. (34)

Military events, damage to bank assets, and changes in relative supplies of the two assets are the leading candidates to explain variations in the relative price of bank notes. (35) Military events are our sole independent variables. We control, as best we can, for damage to bank assets by using the price of North Carolina bank notes. After the Proclamation, the supply of bank notes did not change by much, if any at all, and the money supply data in the Confederacy is not reported frequently enough to be useful for our purposes. (36) However, we check our results for robustness by including dummy variables to control for major financial events.

Entering separate dummies for each battle allows the different engagements to be of varying importance and to differ in sign from what military histories suggest. A Northern victory on the battlefield raises the probability of an ultimate Southern defeat in the war and so advances the time of that defeat and the time of the demise of Confederate currency. On the other hand, a Southern victory augurs well for the Confederate cause and the future value of Confederate currency. As a result, Northern victories should raise the Confederate dollar price of North Carolina notes and Southern victories should cause these bank notes to depreciate against Confederate currency.

The results are shown in Table 1. The pattern we see in the coefficients is broadly consistent with our expectation that Northern victories typically led to an appreciation of bank notes against the Confederate dollar, while Southern victories caused bank notes to depreciate. Clear and dramatic Northern victories at Gettysburg and Atlanta have positive, large, significant coefficients. These Northern triumphs, along with the North's victories at Chattanooga; Early's retreat from Washington, DC; Sherman's March; and the fall of Savannah are associated with significant increases in the growth rate of the Confederate dollar price of North Carolina bank notes. The tactical defeats of Northern forces at the battles of the Wilderness and Spotsylvania and at Cold Harbor are reflected in large, negative, significant coefficients.

The coefficients on Chickamauga and Mobile Bay are of the same magnitude but have the opposite sign of the constant. Neither the constant nor the coefficients on these battles are significant. This pattern occurs because the price of North Carolina bank notes did not change after these battles, and these coefficients reproduce that outcome. Deleting these battles from the regression does not substantively change the size or significance of the coefficients on other battles. In short, the market for North Carolina notes did not respond to these battles.

All of the significant coefficients are of the expected sign, with the exception of the Battle of Stones River. At Stones River in Murfreesboro, Tennessee, the Northern army, commanded by General Rosencrans, squared off against the Confederates led by General Bragg. Bragg attacked and staggered the Northern army, but Rosencrans was able to hold his position, and Bragg eventually retreated. His early successes in this battle led Bragg to telegraph Richmond on December 31 with news of a victory. (37) The Charleston Mercury, on the basis of a report from the Chattanooga Rebel, still reported this battle as a Southern victory in its January 6th edition. It may have been this incorrect information that confused Richmond markets.

The battles of Mobile Bay and Chancellorsville also have coefficients that are not of the expected sign, but these estimates are not significant. There is reason to believe that other news may have confounded the implications of these battles. The Battle of Mobile Bay is a Northern triumph, but has a negative and insignificant coefficient. This battle was a naval engagement that occurred on August 5, 1864. However, Mobile Bay was not the only war news of the week. Six days earlier, on July 30, the North had attempted to break the siege of Petersburg by tunneling under the Confederate works and setting off a tremendous explosion that created an enormous crater and collapsed the Confederate defense. However, the North failed to capitalize on the opportunity and instead met with a startling defeat. The Battle of the Crater was a dramatic Northern setback and took place near the financial markets from which we obtained our data. We conjecture that the coefficient on the Battle of Mobile Bay also reflects the effects of the Battle of the Crater, a Northern battlefield failure. Chancellorsville was a Southern victory on the battlefield, but its effect may have been nullified by the death of the legendary Confederate General Stonewall Jackson, who was killed by "friendly fire." The estimated effect for Fredricksburg, Chickamauga, and the last leg of Sherman's March are statistically insignificant. Because the March was more or less continuous, markets may have already anticipated its final implications. The failure of both Fredricksburg and Chickamauga to have significant effects is somewhat more puzzling but may just be the result of noise in our data.

We checked the robustness of the effect of war events to the inclusion of dummy variables that indicate important financial legislation. We consider the National Bank Act, as well as the first and third funding acts of the Confederacy. McCandless (1996) suggested that the National Bank Act was important because it eased the financing burden of the North. (38) In addition, the National Bank Act placed a strict limit on the number of national bank notes to be issued. This note limitation created the potential of substantial seigniorage on the new currency to be shared by the new banks, northern bond holders, and the U.S. Treasury. The people of the Confederacy and much of the American West would not share the seigniorage. (39) The National Bank Act thus provided a brilliant coup for the Northern war effort. It raised revenue, shored up political support among Northern financial circles to pursue a full military victory, and attacked the purchasing power of the Confederate printing presses. We would thus expect this legislation to cause bank notes to appreciate against Confederate currency.

The South passed several pieces of financial legislation designed to encourage the conversion of currency into bonds. (40) At the beginning of the war, Confederate currency holders held an implicit option to convert their currency into 8% government bonds at par. On three occasions, legislation passed that changed this option. The first funding act set April 22, 1863, as the deadline for converting currency into bonds. After that date, any currency issued before December 1, 1862, would only be convertible into 7% bonds. (41) As the deadline neared, the demand for old notes rose relative to new ones and old notes began to sell at a premium. If this increase in the demand for old notes came, at least in part, from the demand for bank notes, the Confederate dollar price of bank notes would likely have fallen. A similar funding law passed in March 1863 set August 1, 1863, as the deadline to convert notes into bonds. Because this funding law went into effect shortly after the Southern debacles of early July, we can't investigate this law's effect. The third and final funding act was passed in February of 1864 and set an April 1 deadline for conversion. The final funding act reduced the supply of Confederate currency by requiring that three dollars in outstanding notes be exchanged for two dollars in new issue, but certain denominations were subject to additional penalties and the execution of the currency exchange proved to be difficult. (42) The supply of outstanding bank notes was not affected by the Currency Reform Act. (43)

Inclusion of the financial dummies does not change our conclusions with respect to war events. (44) Those battles that were significant without the financial dummies are also significant with them. Similarly, those battles that were not significant remain so when financial dummies are removed. Two of the three financial variables are themselves large and significant. As expected, the National Bank Act raised the premium on bank notes and the third funding act sharply reduced it. The coefficient on the first funding act has the expected negative sign but is not statistically significant.

7. Conclusion

Privately issued bank notes and Confederate currency both circulated in the South during the Civil War. In a simple model of the demand for two monies, we show that, if both currencies are expected to trade at par in the future, then they will trade at par in the present. Early in the Civil War, Lincoln's mission was to preserve the Union. For almost the first year and a half of the Civil War, the South could have laid down its arms and returned to the Union with its antebellum social and political structure intact. Lincoln's policies had led the South to believe that it could return "unhurt" no matter how long the rebellion persisted. From October of 1861 through October of 1862, bank notes and Confederate currency did trade at par. We argue that the tranquility in the market for bank notes reflected Southern expectations that, one way or another, the conflict would end with their social structure intact.

By the summer of 1862, Lincoln recognized that he had to raise the stakes of the war and he did so by issuing the Emancipation Proclamation. The Proclamation forever freed the slaves in the rebelling states and thus committed the North to changing the way of life in the South. The market for Southern bank notes clearly reflects this change in mission from reunification to restructuring and occupation. After the elections of 1862, bank notes traded at premiums against the Confederate dollar and the magnitude of these premiums responded to the outcome of battles. The market for bank notes thus reflected the real possibility that, after the Proclamation, a Southern defeat would lead not only to the destruction of the South's social system, but also to the destruction of its currency.

Table 1. North Carolina State Bank Note Prices
and War Events after the Elections of 1862

                                                       Coefficient
Variable                                                (t-Star)

Constant                                                  0.16
                                                         (1.69)
December 1862: Fredricksburg                              0.10
                                                         (1.04)
May 1863: Chancelorsville                                 0.16
                                                         (1.68)
September 1863: Chickamauga                              -0.16
                                                         (1.69)
May 1864: Wilderness and Spotsylvania                    -0.92
                                                         (9.60)
June 1864: Cold Harbor                                   -0.84
                                                         (8.78)
January 1863: Stones River*                              -3.2
                                                         (3.38)
July 1863: Gettysburg and Vicksburg*                      1.30
                                                        (13.61)
November 1863: Chattanooga*                               0.41
                                                         (4.30)
July 1864: Early's Retreat*                               0.46
                                                         (4.82)
August 1864: Battle of Mobile Bay*                       -0.16
                                                         (1.69)
September 1864: Fall of Atlanta*                          0.43
                                                         (4.48)
November 1864: Sherman's March*                           1.15
                                                        (12.0)
Sherman's March Continues*                                1.39
                                                        (14.6)
Sherman's March Continues and Hood's Army Destroyed*      0.14
                                                         (1.41)
December 1864: Savannah Falls*                            0.42
                                                         (4.43)

The dependent variable is the average daily rate of growth
of the Confederate dollar price of North Carolina state
bank notes expressed as a percentage. The sample begins
in November 1862 and ends in February of 1865. There are
three missing observations.  The r-statistics are in
parentheses and are based on Newey--West standard errors.
Events in bold are typically taken as Northern victories,
while the others are typically thought of as Southern
triumphs. The adjusted [R.sup.2] is 0.13.

Note: Events are typically taken as
Northern victories, indicate with *.

We would like to thank Richard Burdekin, David Hakes, William Hutchinson, Jay Marchand, Ken McCormick, Janet Rives, Clifford Thies, Marc Weidenmier, and two anonymous referees for very helpful comments, and we would especially like to thank Schuyler Porche for invaluable assistance in collecting the data.

(1) McPherson (1989, p. 503).

(2) For a discussion of Lincoln's views on slavery and the evolution of his war policy, see McPherson (1996, chapter 13).

(3) See McPherson (1989, p. 558).

(4) Pecquet (1995) noted the divergence of bank notes from par and attributed it to the National Banking Act and the Emancipation Proclamation. Pecquet dated the break in early January of 1863. With our new. more frequent, and more precise data series, it is clear that bank notes diverged from par two months earlier on November 5. 1862. This new timing leads to a change in emphasis, away from the National Bank Act and toward the 1862 Congressional elections, which were viewed as a referendum on the Emancipation Proclamation. We also differ from Pecquet in that we develop a theoretical model of the demand for two currencies and use regression analysis to examine the effects of war events on the price of bank notes after the 1862 Congressional elections.

(5) The relation between the Confederate money supply and inflation has been explored by Lerner (1955, 1956), Godfrey (1978), and Burdekin and Weidenmier (2001). The role of fiscal policy and currency reforms has been examined by Lerner (1954), Pecquet (1987), Burdekin and Langdana (1993), Grossman and Han (1996). and Burdekin and Weidenmier (2003). The behavior of financial markets during the conflict has been studied by Roll (1972), Calomiris (1988), Davis and Pecquet (1990), Weidenmier (2000), and Brown and Burdekin (2000). The use of interest-beating currency has been the topic of Maiken and Woodward (1999) and Burdekin and Weidenmier (2002). McCandless (1996) and Willard, Guiimane, and Rosen (1996) identify war events by looking at the gold price of currency.

(6) Throughout the war, the Confederacy raised 62% and 20% of its revenues through currency issues and bond sales, respectively. Taxes and miscellaneous sources made up the rest. See Godfrey (1978, p. 14).

(7) Florida was in the process of reestablishing its banking system in 1860 after several years of prohibition (see Johns 1963). See also Rockoff (1975, pp. 125-30) and Schweikart (1987, p. 169).

(8) For analyses of free-banking laws and their operation, see Rolnick and Weber (1983), Rockoff (1975, pp. 125-30), and Schweikart (1987, p. 169).

(9) The Southern states encouraged and later authorized their banks to suspend specie payments in exchange for emergency "war loans" needed to make hasty war preparations even before the Confederacy could be organized. For a discussion of the 1860-1861 Southern banking crisis, see Schweikart (1985; 1987, pp. 284-96), Godfrey (1978, p. 69), and Schwab (1901, pp. 124-31).

(10) As we would expect, the banks that suspended convertibility saw their notes trade at discounts below gold in December 1860 (before the issue of any Confederate currency). The notes from Northern Alabama banks traded at 2% below par. Georgia and South Carolina notes traded at 3% to 6% below par. and the North Carolina and Tennessee banks traded at 5% to 10% below par (Mobile-Price Current 1860).

(11) Bragg (1941, pp. 67-72).

(12) In a postwar petition to the state legislature, eight Georgia bankers complained: "In the year 1860 the act (suspending specie payments) was continued in force, and again reenacted in 1861, but with a proviso, that a redemption of their bills should be made in the Treasury Notes of the state of Georgia or the Confederate States at par. Prior to this proviso, the banks loaned a large amount of bills to the state of Georgia, and subsequently were compelled to obtain the Treasury Notes of the Confederate States, and the State of Georgia Notes to comply with the proviso which by the legislature was made a condition of their existence. By an act of the legislature assented to on the 29th of November, 1862, relief (specie suspension) was again extended to the banks, but upon the condition that said banks receive the Treasury Notes of the state of Georgia and of the Confederate Slates in payment of all dues and also upon deposit at par value" (Savannah, Georgia Banks 1865 Georgia State Archives, Atlanta, Georgia).

(13) Pickens (1862, pp. 8-10).

(14) The tax codes were rewritten by every Confederate state during 1861 or early 1862. See Morgan (1985).

(15) Alabama Governor correspondence. Thomas N. Watts Papers (Alabama State Archives, Montgomery, Alabama).

(16) To avoid confusion, there is a distinction between private bank notes issued under state law and State Treasury notes. For example, Louisiana notes traded at a discount in 1864-1865 (Pecquet 1988, pp. 272-8) while bank notes invariably traded at premiums after October 1862.

(17) Rolnick and Weber (1988) have shown that the demand for bank notes during the flee-banking era depended on the expected value of their backing. We abstract from the backing of bank notes by bank assets and instead focus on the backing of Confederate currency.

(18) Economists argue that the value of currency depends on its backing in general and expectations of fiscal policy in particular. See, for example, Sargent and Wallace (1981). Mitchell (1903) argued that war news during the Civil War altered people's estimation of the prospect of redemption of Confederate currency. Burdekin and Langdana (1993) found that the gold value of Confederate currency was more closely linked to war events than to changes in Confederate financial policies, because war events served as indicators of future fiscal policy and hence the likelihood of redemption.

(19) For a clear and detailed discussion of this approach to the demand for money, see Walsh (1998).

(20) To derive these conditions in this form, divide both sides of the intertemporal budget constraint by [P.sub.t] and then differentiate with respect to [d.sub.t]/[P.sub.t] and [n.sub.t]/[P.sub.t].

(21) Caldwell (1935, pp. 91-6).

(22) Bankers' Magazine (August 1862, p. 96) and Caldwell (1935, pp. 91-6).

(23) Bank of Cape Fear, Washington, Ledger Book; Bank of Cape Fear, Washington, Letter Book; Bank of North Carolina, Windsor Branch, Cashiers Book as cited in Pecquet (1995, p. 142).

(24) See Pecquet (1995, p. 142).

(25) Thian (1880, p. 567).

(26) The banks are typically identified as the "principal banks of Georgia, S.C., N.C. and Alabama." See. for example, the Richmond Examiner, December 12, 1861. The data were collected from various issues of the Richmond Daily Whig, Richmond Enquirer, Richmond Examiner, and Richmond Sentinel from October of 1861 to April of 1864. Beginning in April of 1864, the data come from the Wilmington Journal, with the exception of the observation on September 6, which comes from the Examiner.

(27) The paper also reported that the Governor's commission to examine the bank had found it solvent and doing a legal business. The reputation of the Pittsylvania bank was evidently restored as its notes were later taken on deposit.

(28) On these points, see McPherson (1996, chapters 8 and 9).

(29) Prices were usually reported without references to volume of transactions, but we have some scattered information on the volume trade at the brokerages. For example, on January 6, 1863, the Richmond Daily Whig reported that trades were in denominations of $5 in multiples of 5. Other examples include a December 12, 1863, report in both the Whig and the Examiner of a trade of $2000 ($6500 CSA paper), a trade of $3000 ($9000 CSA paper) worth of Virginia notes in March of 1864 cited by the Whig and a trade of $400 ($1540 CSA paper) of Georgia notes reported in December of 1864 also by the Whig. It may also be worth noting that, when papers omitted quotes, they often said that "nothing or little was doing" in the market. Thus, it seems clear that the price quotes we observe are quotes from significant, actual trades near or on the reporting date.

(30) The Democrats did propose a resolution on December 5 that "anyone attempting to turn the war into one that would overthrow the institutions of the states shall be guilty of a high crime." but, of course, the resolution failed.

(31) The value of bank notes depended not only on the fate of the Confederacy, but also on the property that backed bank notes. Campaigns and battles stretched across Virginia throughout the conflict, but North Carolina property was in large measure spared from the ravages of war;, and so North Carolina bank note prices would presumably be relatively free of variations caused by property destruction. However, the results described below are not meaningfully changed when we use Virginia bank note prices. The only difference of note is that, in the regression with Virginia note prices, the coefficient on the fall of Atlanta is roughly half the size of its North Carolinian counterpart. Full results using Virginia bank note prices may be found at www.cba.uni.edu/kanago/data.htm.

(32) Our data on North Carolina and Virginia bank note prices may be found at www.cba.uni.edu/kanago/data.htm.

(33) The dates of the battles and the associated observation for the dummy variables may be found at www.cba.uni.edu/kanago/data.htm.

(34) See Weidenmier (2002) for a discussion of the spread of information, particularly with respect to news about Gettysburg.

(35) Burdekin and Langdana (1993, p. 353) found that "... Fiscal policy variables cannot satisfactorily account for the fluctuations in the Confederate price level. Rather, we find 'war news' to be of key importance in determining price movements...."

(36) See Godfrey (1978, pp. 66-7).

(37) See McPherson (1989, p. 582).

(38) The National Bank Act was originally called the National Currency Act. The National Currency Act became known as the National Bank Act in 1864 when it was revised.

(39) Henry C. Carey, a contemporary economist who generally supported Republican policies, such as protective tariffs, condemned the National Bank Act as monopolistic because it restricted note issue. See Helderman (1931, pp. 146-7).

(40) See Burdekin and Weidenmier (2003) for an analysis of these funding acts and their asymmetric effect on asset and commodity prices.

(41) Notes issued after December 1, 1862, could only be converted into 7% bonds.

(42) For a detailed description of the February 1864 Currency Reform Act and an analysis of its effects in the trans-Mississippi, see Pecquet (1987). For an analysis of the relationship between the supply of Confederate money and prices in two isolated sections within the Confederacy, see Burdekin and Weidenmier (2003).

(43) Godfrey (1978. p. 75).

(44) These results may be found at www.cba.uni.edu/kanago/data.htm.

References

Bankers' Magazine and Statistical Register, August 1862, p. 96.

Bragg, Jefferson Davis. 1941. Louisiana in the Confederacy. Baton Rouge, LA: L.S.U. Press.

Brown, W. D., and Richard C. Burdekin. 2000. Turning points in the U.S. Civil War: A British perspective. The Journal of Economic History 60:216-31.

Burdekin, Richard C. K., and Farrokh K. Langdana. 1993. War finance in the Southern Confederacy, 1861-1865. Explorations in Economic History 30:352-76.

Burdekin, Richard C. K., and Marc D. Weidenmier. 2001. Inflation is always and everywhere a monetary phenomenon: Richmond vs. Houston in 1864. American Economic Review 91:1621-30.

Burdekin, Richard C. K., and Marc D. Weidenmier. 2002. Interest-beating currency and legal restrictions theory: Lessons from the Southern Confederacy. Cato Journal 22:199-209.

Burdekin, Richard C. K., and Marc D. Weidenmier. 2003. Suppressing asset price inflation: The Confederate experience, 1861-1865. Economic Inquiry 41:420-32.

Caldwell, Stephen. 1935. A banking history of Louisiana Baton Rouge. Baton Rouge, LA: L.S.U. Press.

Calomiris, Charles. 1988. Price and exchange rate determination during the greenback suspension, Oxford Economic Papers 40:715-50.

Catton, Bruce. 1971. The Civil War. New York, NY: Fairfax Press.

Charleston Mercury. 1863. The Battle of Murfreesboro. 6 January.

Davis, George K., and Gary M. Pecquet. 1990. Interest rates in the Civil War South. Journal of Economic History 50:133-48.

Godfrey, John M. 1978. Monetary expansion in the Confederacy. New York: Arno Press.

Grossman, H. I., and T. J. Han. 1996. War debt, moral hazard, and the financing of the Confederacy. Journal of Money, Credit and Banking 28:200-15.

Helderman, Leonard C. 1931. National and state banks: A study of their origins. Boston and New York: Houghton and Mifflin Co.

Johns, John E. 1963. Florida during the Civil War. Gainsville, FL: The University of Florida Press.

Lerner, Eugene M. 1954. The monetary and fiscal programs of the Confederate government, 1861-1865. Journal of Political Economy 62:506-22.

Lerner, Eugene M. 1955. Money prices and wages in the Confederacy, 1861-1865. Journal of Political Economy 63:20-40.

Lerner, Eugene M. 1956. Inflation in the Confederacy, 1861-65. In Studies in the quantity theory of money, edited by Milton Friedman. Chicago: University of Chicago Press, pp. 163-75.

Maikinen, G. E., and G. T. Woodward. 1999. The use of interest beating currency in the Civil War: The experience below the Mason-Dixon Line. Journal of Money, Credit and Banking, 31:121-9.

McCandless, George T., Jr. 1996. Money, expectations, and the U.S. Civil War. American Economic Review 86:661-71.

McPherson, James M. 1989, Battle cry of freedom: The Civil War era. New York: Ballantine Books.

McPherson, James M. 1996, Drawn with the sword. New York: Oxford University Press.

Mitchell, Wesley C. 1903. A history of greenbacks. Chicago: The University of Chicago Press.

Mobile Price Current. 1860. 15 December.

Morgan, James F. 1985. Graybacks and gold: Confederate monetary policy. Pensacola, FL: Perdido Press.

The New York Herald. 1863. The situation. 5 November, p. 6.

Pecquet, Gary M. 1987. Money in the trans-Mississippi Confederacy and the Confederate Currency Reform Act of 1864. Explorations in Economic History 24:218-43.

Pecquet, Gary M. 1988. Public finance in Confederate Louisiana. Louisiana History 29:253-98.

Pecquet, Gary M. 1995. Southern banking during the Civil War: A Confederate tool and Union target. In Money and banking: The American experience. Fairfax, VA: George Mason University Press, pp. 133-62.

Pickens, Governor F. W. 1862. Message No. 1 of His Excellency F.W. Pickens to the Regular Session of November 1862. Columbia, South Carolina: State of S. C. Press (Columbia, SC: South Carolina Library, University of South Carolina).

Pratt, Fletcher. 1948. Ordeal by fire. New York: William Sloane Associates, Inc.

Richmond Daily Whig. 5 November 1862-12 December 1864.

Richmond Enquirer. 1 October 1861-31 October 1862.

Richmond Examiner. 12 December 1861-31 September 1864.

Richmond Sentinel. 1 October 1861-30 April 1864.

Rockoff, Hugh, 1975. The free banking eta: A reexamination. New York: Arno Press.

Roll, Richard. 1972. Interest rates and price expectations during the Civil War. Journal of Economic History 32:476-98.

Rolnick, Arthur, and Warren E. Weber. 1983. New evidence on the free banking era. American Economic Review 73:1080-91.

Rolnick, Arthur, and Warren E. Weber. 1988. Explaining the demand for free bank notes. Journal of Monetary' Economies 21:47-71.

Sargent, Thomas J., and Neil Wallace. 1985. Some unpleasant monetarist arithmetic. Quarterly Review Federal Reserve Bank of Minneapolis 9:1-17.

Schwab, John C. 1901. The Confederate States of American, 1861-1865: A financial and industrial history of the Civil War. New York: Schribner and Sons.

Schweikart, Larry. 1985. Secession and Southern Banks. Civil War History 31:111-25.

Schweikart, Larry. 1987. Banking in the American South from the age of Jackson to Reconstruction. Baton Rouge, LA: L.S.U. Press.

Thian, Raephael P. (compiler). 1880. Correspondence of the Treasury Department of the Confederate States of America: Appendix Part 4. Washington, DC.

Walsh, Carl E. 1998. Monetary theory and policy. Cambridge, MA: The MIT Press.

Weidenmier, Marc D. 2000. The market for Confederate bonds. Explorations in Economic' History 37:76-97.

Weidenmier, Marc D. 2002. Turning points in the U.S. Civil War: Views from the grayback market. Southern Economic Journal 68:875-90.

Willard, Kristen L., Timothy W. Guinnane, and Harvey S. Rosen. 1996. Turning points in the Civil War: Views from the greenback market. American Economic Review 86:1001-18.

Wilmington (North Carolina) Journal. April 1864-February 1865.

Received November 2002; accepted March 2003.

Gary Pecquet, * George Davis, ([dagger]) and Bryce Kanago ([double dagger])

* Tulane University, New Orleans, LA 70118; E-mail [e-mail].

([dagger]) Department of Economics, Miami University, Oxford, OH 45056, USA; E-mail [e-mail]; corresponding author.

([double dagger]) Department of Economics. University of Northern Iowa, Cedar Falls, IA 50614, USA; E-mail [e-mail].